Maximum Bulk Density of British Columbia
Forest Soils from the Proctor Test: Relationships
with Selected Physical and Chemical Properties
Yihai Zhao
Maja Krzic*
Faculty of Forestry
Univ. of British Columbia
Vancouver, BC V6T 1Z4
Canada
Chuck E. Bulmer
British Columbia Ministry of Forests and Range
Research Branch
Vernon, BC V1B 2C7
Canada
Margaret G. Schmidt
FOREST, RANGE & WILDLAND SOILS
Dep. of Geography
Simon Fraser Univ.
Burnaby, BC V5A 1S6
Canada
The widespread use of heavy equipment during timber harvesting and site preparation can lead
to reduced soil productivity and warrants development of new methods to assess compaction. We
evaluated the effects of soil particle density, organic matter, particle size distribution, extractable
oxides, and plastic and liquid limits on the maximum bulk density (MBD) of forest soils in British
Columbia. Soil samples were collected from 33 sites throughout British Columbia, covering the
major forest and soil types of the province. The standard Proctor test was used to determine
MBD and related parameters, including the gravimetric water content (WMBD) and porosity
(fMBD) at which MBD was achieved. The significance levels of single soil properties in predicting
MBD were in the order plastic and liquid limits, organic matter, oxalate-extractable oxides, and
particle size distribution. For all samples, liquid limit and clay were most closely related to MBD
(R2 = 0.83). Addition of organic matter to the model increased the regression coefficients, and
oxidizable organic matter caused a greater increase than did total C. Stratification of the sample
set into groups based on plasticity led to higher R2 values in multiple regressions, and different
soil properties were important for nonplastic soils than for those with high, moderate, and low
plasticity. Prediction with multiple regression explained the most variation in MBD for nonplastic
soils, while properties of highly plastic soils explained the least variation in MBD and moderately
plastic soils were intermediate. Based on our findings, we propose an approach for using MBD to
help better interpret bulk density data in forest soil compaction studies.
Abbreviations: BWBS, Boreal White and Black Spruce biogeoclimatic zone; CDF, Coastal Douglas-fir
biogeoclimatic zone; CWH, Coastal Western Hemlock biogeoclimatic zone; f, porosity; fMBD, porosity
at MBD; ICH, Interior Cedar–Hemlock biogeoclimatic zone; IDF, Interior Douglas-fir biogeoclimatic zone;
LTSP, Long-Term Soil Productivity Study; MBD, maximum bulk density; PCA, principal component analysis;
SBS, Sub-Boreal Spruce biogeoclimatic zone; gravimetric water content at which MBD was achieved.
M
echanized forest harvesting operations apply heavy
weights to soil, which often leads to compaction.
Reduced tree volume and height growth caused by compaction
have been reported in various parts of North America (Wert
and Thomas, 1981; Page-Dumroese et al., 1998), and it can
take decades (as long as 70 yr) for compacted soils to naturally recover to their predisturbance conditions (Froehlich et al.,
1985; Miller et al., 1996). Compaction is a process of increasing the soil bulk density (and decreasing porosity) by application of mechanical forces to the soil. Successful planning to
minimize compaction depends on knowledge of the distribution of soil types in a given area, coupled with a knowledge
of the behavior of the soils in response to compactive effort.
Many regions of North America and elsewhere have extensive
Soil Sci. Soc. Am. J. 72:442-452
doi:10.2136/sssaj2007.0075
Received 21 Feb. 2007.
*Corresponding author (krzic@interchange.ubc.ca).
© Soil Science Society of America
677 S. Segoe Rd. Madison WI 53711 USA
All rights reserved. No part of this periodical may be reproduced or
transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher.
Permission for printing and for reprinting the material contained
herein has been obtained by the publisher.
442
soil resource inventories, but work on the site-specific effects of
compaction needs to be better developed.
To maintain sustainable soil productivity, it is necessary
to assess the ability of a soil to support plant growth after
machines have traveled over it. Soil scientists studying sustainability have traditionally measured bulk density as an indicator of compaction, and this measurement is also made because
bulk density is a key soil property for determining site nutrient
contents. Despite this, limiting values for bulk density have
not been defined for the wide range of soil conditions typical in forests, primarily because such limiting values are different for soils with varying texture, organic matter content, and
other properties. Establishment of limiting values would be
beneficial for soil scientists and land managers.
One approach to better evaluate the state of soil compaction among soil types involves expressing the actual bulk density as a percentage of some reference compaction state (Lipiec
et al., 1991; Topp et al., 1997; Lipiec and Hatano, 2003). The
idea of comparing soil physical conditions at field sites to a reference state was also proposed by Joosse and McBride (2003),
who proposed comparisons based on the void ratio to evaluate
the soil quality of agricultural sites. Such comparisons would
allow conditions from a wide range of soil types to be evaluated
using a single threshold limit, much as the critical limits of soil
mechanical resistance and air-filled porosity appear to be relatively independent of soil type (Hakansson and Lipiec, 2000;
SSSAJ: Volume 72: Number 2 • March–April 2008
Zou et al., 2001). Therefore, use of a reference state could
potentially enhance interpretations in soil compaction studies.
Various parameters for a reference compaction state have
been proposed (Carter, 1990; da Silva et al., 1994; Hakansson
and Lipiec, 2000), but the maximum bulk density (MBD)
determined by the standard Proctor compaction test (ASTM,
2000) is rigorously defined, is readily determined with standard
test equipment, and has been used in several studies (Carter,
1990; Smith et al., 1997; Aragon et al., 2000). The potential
advantages of using MBD as a reference compaction state can
only be realized if the soil samples used to determine it reliably
represent site conditions, and this can create challenges in forest
soils. Unlike agricultural soils, where soil type is often relatively
consistent within a particular field, the properties of forest soils
are known to vary widely across short distances (Courtin et
al., 1983) on many forested sites in response to more variable
topography and the absence of tillage to mix and homogenize
surface layers. Such variation would require a large number of
samples to be taken to determine MBD, and some alternative
method to predict MBD would be beneficial.
The standard Proctor method (ASTM, 2000) evolved from
studies by civil engineers (Proctor, 1933) on the compaction of
soils for dam and road foundations. Two parameters are obtained
from this method: MBD and the critical water content at which
MBD is achieved for a given amount of energy (WMBD). The
compactive force applied in the Proctor test as it is used in engineering studies has evolved over the years to make it more applicable to changing needs. Despite this, no information is currently
available to determine whether different levels of applied force
would improve interpretations of compaction effects on growth
(Hakansson and Lipiec, 2000). Therefore the standard Proctor test
is commonly used in productivity studies (Carter, 1990).
The variation in MBD as determined by the standard
Proctor test for a range of soils has been attributed to changes
in soil organic matter, particle size distribution, Fe and Al
oxides, or plastic and liquid limits. For example, quantity as
well as quality of organic matter has been determined to have
effects on MBD (Soane, 1990; Aragon et al., 2000), and both
organic C (Donkin, 1991; Smith et al., 1997; Krzic et al.,
2004) and readily oxidizable organic matter (Ball et al., 2000)
have been used to predict MBD. Cementing agents, such as
Fe, Al, or Mn oxides (in acidic soils) and carbonates (in calcareous soils) enhance aggregate stability, contributing to high
soil shear strength (Yee and Harr, 1977). Dorel et al. (2000)
reported that Caribbean Andosols and Nitisols (FAO, 1998)
or Andisols and Alfisols (according to Soil Survey Staff, 2006)
were more resistant to compaction because of the presence
of stable microaggregates containing halloysite and Fe oxide.
Larson et al. (1980) found that among 36 agricultural soils
from around the world, soils with predominantly kaolinite or
Fe oxide in the clay fraction had lower MBD than soils with
predominantly 2:1 type clays.
The MBD was significantly correlated with clay, fine silt,
coarse silt, medium sand, and fine sand, and the clay + silt fraction
had the strongest (inverse) correlation with MBD (R = 0.79) on 26
South African forest soils (Smith et al., 1997), while Nhantumbo
and Cambule (2006) also showed a relationship between clay content and MBD. Peakedness (kurtosis) and symmetry (skewness)
of the particle size distribution curve have also been suggested as
SSSAJ: Volume 72: Number 2 • March–April 2008
parameters in predicting MBD (Webster and Oliver, 1990). Wellgraded soils, as indicated by a low coefficient of kurtosis, tend to
have a higher MBD. A linear relationship between MBD and kurtosis (R2 = 0.82) was reported by Moolman (1981), while Smith
et al. (1997) showed that MBD decreased as the degree of kurtosis
increased, but the relationship was not strong (R = −0.48).
Particle density of mineral soils dominated by Fe oxides
and heavy minerals can range from 3.0 to 5.0 Mg m−3
(Padmanabhan and Mermut, 1995; Ruhlmann et al., 2006),
while organic soil may have a particle density as low as
0.84 Mg m−3 (Redding and Devito, 2006). Since soil bulk
density is influenced by particle density, consideration of particle density should be made when predicting MBD. This is particularly important when the soils examined have a wide variation in particle density, as they may for groups of soils developed
on diverse geologic materials in mountainous terrain.
Due to the complex interrelationships among soil properties, attempts have also been made to combine several soil
properties when predicting MBD. For example, variation in
MBD was predicted well by the liquid limit, organic C, and
sand (R2 = 0.98) in a study performed by Howard et al. (1981)
on 14 forest and rangeland soils in California. Similarly, Ball
et al. (2000) showed that MBD, WMBD, and total porosity at
MBD were predicted (R2 = 0.49, 0.55, and 0.43, respectively)
by a combination of liquid limit and readily oxidizable organic
matter for a range of cultivated soils in Great Britain.
Based on these findings, and as part of a larger study performed throughout British Columbia to determine the effects
of compaction on forest soil productivity and tree growth, we
evaluated the potential for predicting MBD based on properties
that can be determined on samples normally collected during
field evaluations of bulk density on forested sites. Our objectives were to: (i) evaluate the relationships between MBD and
WMBD as determined by the standard Proctor test and other
soil properties for a wide range of British Columbia forest soils;
(ii) identify the soil properties most important for predicting
MBD; and (iii) describe a proposed method for using MBD as
a reference bulk density in forest soil compaction studies.
MATERIALS AND METHODS
Study Sites
A total of 147 soil samples were collected from 33 study sites
(Table 1) located in timber-growing areas within the Boreal White
and Black Spruce (BWBS), Sub-Boreal Spruce (SBS), Interior Douglasfir (IDF), Interior Cedar–Hemlock (ICH), Coastal Douglas-fir (CDF),
and Coastal Western Hemlock (CWH) biogeoclimatic zones of British
Columbia (Meidinger and Pojar, 1991). Ninety-three samples from 16 of
the study sites were included in a previous study by Krzic et al. (2004).
The most common soil textural classes were silt loam and loam,
with substantial variation often occurring within study sites. Soils were
classified as Inceptisols or Brunisols and Gleysols (according to the
Soil Classification Working Group, 1998), Alfisols or Luvisols (Soil
Classification Working Group, 1998), and Spodosols or Podzols (Soil
Classification Working Group, 1998), which covered the range of pedogenetic development in British Columbia. The majority of the soils were
developed on glacial till, with the exception of one Inceptisol in the
ICH developed on colluvium, two Alfisols in the SBS and ICH zones
developed on lacustrine parent material, and seven Inceptisols in the CDF
and CWH zones developed on glaciomarine parent material (Table 1).
443
Table 1. Site description, biogeoclimatic (BEC) zones, and annual precipitation for 33 study
sites throughout British Columbia.
followed by sieving through a 4.75-mm
sieve and further air drying. To carry out
Study site
BEC† Precipitation Soil suborder
Parent material
the test, an initial estimate was made for
mm
each sample of the water content at which
Black Pines
IDF
279
Cryalf
Eolian veneer over glacial till
MBD would be achieved. Because WMBD
Dairy Creek
IDF
279
Cryalf
Eolian veneer over glacial till
is typically slightly less than the plastic
Emily Creek
IDF
424
Cryept
Glacial till
limit, the initial estimate involved deterKiskatinaw
BWBS
482
Cryalf
Glaciofluvial veneer over glacial till
mining the water content of a sample that
Kootenay East
IDF
424
Cryept
Glacial till
had been moistened to the point where,
Log Lake
SBS
615
Udept
Glacial till
after squeezing in the hand, it would
McPhee Creek
ICH
755
Udept
Colluvium
remain
in a lump when hand pressure was
Mud Creek
IDF
424
Cryept
Glacial till
released,
but would break cleanly into two
O’Connor Lake
IDF
279
Cryalf
Eolian veneer over glacial till
pieces when “bent” (ASTM, 2000). Water
Rover Creek
ICH
755
Udept
Colluvium
was then added to a 2.3-kg subsample until
Skulow Lake
SBS
425
Cryalf
Glacial till
Topley
SBS
530
Cryalf
Glacial till
it reached the estimated water content, and
Aleza Lake
SBS
930
Cryalf
Lacustrine
then four more subsamples were prepared,
Miriam Creek
ICH
420
Cryalf
Glacial till
two with soil water content (W) below,
Vama Vama
ICH
601
Cryalf
Lacustrine
and two with W above this value. The five
Gates Creek
ICH
410
Cryalf
Glacial till
subsamples were then left in sealed plastic
Phoenix
ICH
450
Cryoll
Glacial till
bags to equilibrate overnight. During the
Aitken
BWBS
464
Cryalf
Glacial till
test, soil was compacted in a standard mold
Bernadet
BWBS
498
Cryalf
Glacial till
(9.43
× 10−4 m3) using a 2.5-kg rammer
Blackhawk
BWBS
619
Cryalf
Glacial till
falling freely from a height of 0.3 m. The
Blueberry
BWBS
489
Cryalf
Glacial till
soil was added to the mold in three layers
Boot Lake
BWBS
581
Cryalf
Glacial till
and 25 blows of the rammer were applied
Apollo
SBS
497
Cryalf
Glacial till
John Prince
SBS
565
Udept
Glacial till
to each layer. Total compactive effort
Weedon
SBS
606
Udept
Glacial till
applied to the sample was approximately
Younges
SBS
615
Cryalf
Glacial till
600 kN m m–3 (or 595 kJ m–3). The
Port Alberni
CWH
2116
Udept
Glaciomarine
compacted sample was used to determine
Duncan Eagle
CDF
1039
Udept
Glaciomarine
bulk density and corresponding water conDuncan Keating
CDF
1039
Aquept
Glaciomarine
tent. Soil water content was determined
Duncan Somenos
CDF
1039
Ustept
Marine or Lacustrine
gravimetrically (w/w) by drying samples
Kennedy Lake
CWH
3295
Udept
Glaciomarine
at 105°C for 16 h. Dry bulk densities vs.
Saanich Cowichan
CDF
906
Aquept
Glaciomarine
W
values were plotted on a graph and the
Saanich Fairbridge
CDF
906
Udept
Glaciomarine
points were fitted with a best-fit curve
† Biogeoclimatic zone: IDF, Interior Douglas-fir; BWBS, Boreal White and Black Spruce; ICH, Interior Cedar–Hemlock; SBS, Sub-Boreal Spruce; CWH, Coastal Western Hemlock; CDF, Coastal Douglas-fir. (third-order polynomial). From the resulting compaction curve, MBD was deterThe sites sampled for this study included 12 long-term soil
mined from either (i) the peak of the curve,
productivity study (LTSP) installations, three long-term landing
or (ii) the highest sample value when the peak of the curve lay below
rehabilitation trials, seven provincial park sites, five oil-explorationthat level. Approximately 0.5 kg from each sample was sieved through
disturbed sites, four road rehabilitation sites, and two stumping-disa 2-mm sieve to determine the percentage of fine fraction, which was
turbed sites (Fig. 1). The LTSP sites in British Columbia are part of the
used to correct MBD. The volume of mineral coarse fragments was
determined from dry mass and assumed to have a particle density
North American LTSP network that includes the USDA Forest Service,
of 2.65 Mg m−3. Fine-fraction MBD was calculated as the mass of
Canadian Forest Service, British Columbia Ministry of Forests and
Range, and various universities and industry groups (Powers, 2006).
dry, coarse-fragment-free mineral soil per volume of moist soil, where
volume was also calculated on a coarse-fragment-free basis. All MBD
Sample locations were selected to be representative of typical site
values are reported on a fine-fraction basis.
conditions. For the LTSP sites, sample locations had been harvested
with minimal soil disturbance. Soils from landing and oilfield rehabilitation trials usually experienced some scalping of surface layers, while
Particle Density
soils from the stumping trials were characterized by some mixing of
Particle density was determined by the gas displacement method
surface soil layers. Samples (?35 kg each) were collected at 0- to 0.1-, 0.1(Flint and Flint, 2002), which was modified so that the expansion
to 0.2-, or 0- to 0.2-m depth after removal of the forest floor (if present).
chamber (instead of the sample chamber) was pressurized to 239 kPa
The number of samples collected at each site varied between two and 12.
with room air. The expansion chamber was then opened to the sample
chamber. Volumes of the two chambers, hose, and transducer were
not measured directly; instead, a model describing the volume–presSoil Analysis
sure relationship was derived based on the changing volume of the
Maximum Bulk Density
sample chamber with known-volume plastic disks. Soil samples <2
The MBD and WMBD were determined using the standard
engineering Proctor test (Proctor, 1933; ASTM, 2000). Soil samples
mm, oven dried at 60°C for 48 h, were used for the test. The samples
used for determination of particle density were not treated to remove
were air dried until friable and then passed through a 19-mm sieve
444
SSSAJ: Volume 72: Number 2 • March–April 2008
organic matter, so the reported particle density for each sample reflects an average particle
density for the entire fine fraction, including
organic and mineral components.
Soil Organic Matter
Soil total C was determined by the dry
combustion method (Nelson and Sommers,
1996) using a LECO analyzer (LECO Corp.,
St. Joseph, MI) on a sample that had passed
through a 2-mm sieve. The estimate of
the “readily oxidizable organic matter” was
obtained as the weight difference before and
after treatment of the sample with H2O2 and
expressed as the gravimetric fraction of the
original mass of soil. Hydrogen peroxide treatment involved heating a 50-g sample with 75
mL of H2O2 (30%) and 300 mL of water to
80°C and adding small increments of H2O2
until no further reaction was observed.
Soil Oxides
Soil oxides of Al, Fe, Mn, and Si were Fig. 1. Location of 33 study sites in British Columbia. Sites 1–3, 5, and 6 are oil-exploration-disturbed
extracted by 0.2 mol L−1 acid ammonium
sites; Sites 4, 7, 11, 15–18, and 22–26 are long-term soil productivity study installations; Sites 13,
14, and 19 are long-term landing rehabilitation sites; Sites 20 and 21 are stumping-disturbed sites;
oxalate solution. This method extracts active
Sites 8–10 and 12 are road rehabilitation sites; and Sites 27–33 are provincial park sites.
Al, Fe, Mn, and Si oxides, including a fraction
of oxides bound by organic matter (Loeppert
and Inskeep, 1996). The extracted ions were measured by inductively
was determined using the one-point Casagrande method (McBride,
coupled plasma spectrometer.
2002). The soil water content at which 20 to 30 blows are required to
close a groove along a distance of 13 mm was determined gravimetrically after drying at 105°C for 16 h. The liquid limit was calculated
Particle Size Distribution
Soil particle size distribution was determined by the hydrometer
using the following equation:
method (Gee and Or, 2002). Samples were pretreated with H2O2
0.12
[3]
LL = W (N 25 )
(30%) and heat, while samples from the IDF zone that may have
contained carbonates were also treated with a NaOAc buffer. Particle
where LL is the liquid limit, N is the number of blows, and W is the
gravimetric water content.
size distribution was described using the Canadian System of Soil
Classification (Sheldrick and Wang, 1993) in terms of the percentage of clay (<0.002 mm), fine silt (0.002–0.005 mm), medium silt
Statistical Analysis
(0.005–0.02 mm), coarse silt (0.02–0.05 mm), very fine sand (0.05–
Distribution of the data was summarized by principal component analysis (PCA) using the SAS PRINCOMP procedure (SAS
0.10 mm), fine sand (0.10–0.25 mm), medium sand (0.25–0.50 mm),
coarse sand (0.50–1.00 mm), and very coarse sand (1.00–2.00 mm).
Institute, 1990). Before the PCA, “missing values” in the data matrix
Peakedness (kurtosis) and symmetry (skewness) of the particle
(i.e., 34 samples without plastic limit and three without liquid limit)
size distribution curve were calculated using the equations proposed
were filled by the SAS PRINQUAL procedure with the minimum
by Webster and Oliver (1990):
generalized variance method (SAS Institute, 1990). Simple regression
analyses between dependent (i.e., MBD, WMBD) and independent
[1]
skewness: Y1 = M 3 M 23/ 2
variables (i.e., soil properties) were run. Samples without plastic or
liquid limits were not included in the regression. In addition, multiple
[2]
kurtosis: Y2 = M 4 M 22 − 3
regression analysis was used to select soil properties that were highly
where M2 = (1/n)∑(Xi − μ)2 is the second moment of the distribution
correlated with dependent variables. A stepwise method was selected
about the mean of the observation, M3 = (1/n)∑(Xi − μ)3 is the third
in multiple regression analysis because certain soil properties (e.g.,
moment of the distribution about the mean of the observation, M4 =
plastic and liquid limit, total C, and oxidizable organic matter) may
(1/n)∑(Xi − μ)4 is the fourth moment of the distribution about the
have multiple impacts on soil MBD. Significance levels of the χ2 score
mean of the observation, n is the number of observations, Xi is the ith
for variable entry and stay were set at 0.25 and 0.10, respectively.
observation, and μ is the mean of the observations.
(
)
RESULTS AND DISCUSSION
Plastic and Liquid Limits
The plastic limit was determined as the gravimetric water content at which a soil sample could be rolled by hand into a thread of
3-mm diameter without breaking (McBride, 2002). The liquid limit
SSSAJ: Volume 72: Number 2 • March–April 2008
The minimum, maximum, and mean values of the soil properties are presented in Table 2, while a partial correlation matrix for
the relationship between dependent variables (MBD, WMBD) and
selected soil properties is presented in Table 3.
445
Table 2. Soil properties for 33 study sites in British Columbia (n = 147).
Soil property
Maximum bulk density, Mg m−3
WMBD†, kg kg−1
Particle density, Mg m−3
Total C, g kg−1
Oxidizable organic matter, g kg−1
Clay, g kg−1
Silt, g kg−1
Sand, g kg−1
Fe oxide, %
Mn oxide, %
Al oxide, %
Si oxide, %
Liquid limit, kg kg−1
Plastic limit, kg kg−1
Min.
0.91
0.09
2.33
1.8
2.2
19
94
29
0.10
0.002
0.08
0.03
0.15
0.14
Max.
1.98
0.50
2.97
77.4
76.7
703
728
851
1.19
0.31
1.56
0.66
0.61
0.57
Mean
1.51
0.22
2.66
23.4
28.6
201
450
349
0.51
0.05
0.32
0.11
0.33
0.26
SD
0.23
0.08
0.10
16.5
15.3
136
114
172
0.25
0.04
0.28
0.11
0.11
0.08
† Water content at which maximum bulk density was achieved.
Relationships between Maximum Bulk Density,
Water Content at Maximum Bulk Density, and
Other Soil Properties
Particle Density
Particle density for soils from our study sites ranged from 2.33
to 2.97 Mg m−3 (Table 2), with relatively large variation observed
among the sites (27 out of 33 sites had within-site particle density
differences >5%). The MBD was significantly (R2 = 0.36) positively correlated with particle density (data not shown).
Although compaction effects on forest soil productivity
are more directly related to changes in porosity, f (volume of
pores/total soil volume), than bulk density, it is bulk density
(and MBD) that is most commonly used to describe the forest
Fig. 2. Change of (a) soil bulk density and (b) porosity with water
content (W) in the standard Proctor test. The bottom curve in
(a) is hypothetical, derived by replacing the particle density of
2.97 with 2.33 Mg m−3.
446
soil condition (Smith et al., 1997). For groups of soils where
particle density varies only slightly, bulk density and MBD
can accurately reflect changes in f as they affect plant growth
because the change in MBD is inversely related to the volume
of pores. For groups of soils with a range of particle densities, however, changes in bulk density and MBD will reflect
both differences in the volume of solids and particle densities,
and the compaction state as it affects growth might be better
described by the change in f rather than by the bulk density.
In Fig. 2, we plotted the Proctor test curve of the soil with
the highest particle density (2.97 Mg m−3), then replaced its
particle density with the lowest one (2.33 Mg m−3) and plotted the five points again. The derived MBD of the hypothetical
soil was 28% lower than that of the soil with the highest particle
density (Fig. 2a), even though f remained the same (Fig. 2b).
In forest compaction experiments where the particle density varies widely, determining the f of field soils and relating
it to the f determined at MBD (i.e., fMBD) could serve as an
alternative index of the soil compaction state. This is analagous
to the approach of Joosse and McBride (2003), who proposed
relating the void ratio of structurally intact agricultural soils
to that of remolded soils subjected to slurry consolidation and
uniaxial compression tests (preconsolidation). The use of f and
fMBD could also be advantageous for evaluating compaction
effects on soils with different parent materials, as described by
Smith et al. (1997).
Soil Organic Matter
The MBD was negatively and WMBD positively related to
total C and oxidizable organic matter (Table 3). The MBD is
commonly considered to be linearly related to organic matter
(Soane, 1990; Zhang et al., 1997; Krzic et al., 2004). In this
study, however, for both total C and oxidizable organic matter,
exponential models gave a better description of the relationship
(R2 = 0.70 and 0.64, respectively; data not shown) than linear
models (R2 = 0.65 and 0.61, respectively; data not shown).
Linear models described well the relationships between WMBD
and total C and oxidizable organic matter (R2 = 0.65 and 0.54,
respectively; data not shown).
Greacen and Sands (1980) reported that increased organic
matter in sandy soils under radiata pine (Pinus radiata D. Don)
forests in South Australia was associated with reduced compaction (measured as bulk density) under a given load. In a
study performed by Ball et al. (1989) on Gleysol and Cambisol
(FAO, 1998) or Inceptisol (Soil Survey Staff, 2006) soils from
Scotland, a reduction in MBD of 0.18 g cm−3 per increase of
1% organic C was observed. In our study, total C had a similar
effect on MBD.
Oxidizable organic matter is used as an indicator of the
soil organic matter quality, since H2O2 oxidizes the colloidal,
humified organic matter but not the fibrous residues (Day,
1965). The relationship between MBD and oxidizable organic
matter has been reported in several studies. For example, Soane
(1990) indicated that highly humified material increased soil
aggregate stability and soils with high oxidizable organic matter
tended to be less compacted. Ball et al. (2000) have shown that
oxidizable organic matter explained 63% of variation in MBD
of British agricultural soils.
SSSAJ: Volume 72: Number 2 • March–April 2008
SSSAJ: Volume 72: Number 2 • March–April 2008
*** Significant at P < 0.001.
** Significant at P < 0.01.
† MBD, maximum bulk density; fMBD, porosity at MBD; WMBD, water content at which MBD was achieved; PD, particle density; oxOM, oxidizable organic matter; FSI, fine silt; MSI, medium
silt; FS, fine sand; MS, medium sand; CS, coarse sand; VCS, very coarse sand; LL, liquid limit; PL, plastic limit. Other variables that were not significantly correlated to MBD, fMBD, or
WMBD are not shown.
1.00
0.90 *** 1.00
1.00
0.36 ***
0.42 ***
1.00
0.39 ***
0.43 ***
0.69 ***
1.00
−0.49 *** 1.00
−0.37 *** 0.77 *** 1.00
−0.34 *** 0.68 *** 0.77 *** 1.00
−0.03
0.31 **
0.58 *** 0.70 *** 1.00
0.11
0.13 *** 0.08
0.20 **
0.14
−0.02
−0.29 *** −0.23 ** −0.19 ** −0.22 *
0.12
−0.32 *** −0.34 *** −0.24 ** −0.12
0.14
−0.16
−0.16 * −0.04
−0.08
1.00
0.13 *
−0.62 ***
−0.53 ***
−0.56 ***
−0.42 ***
−0.16 *
0.24 **
0.29 ***
0.18 *
1.00
0.70 ***
−0.26 *
−0.49 ***
−0.50 ***
−0.57
−0.58 ***
−0.22 **
0.26 ***
0.42 ***
0.21
1.000
0.20 **
0.25 **
0.23 **
−0.24 **
−0.21 **
−0.16 *
−0.03
0.26 **
0.46 ***
0.69 ***
0.73 ***
1.00
0.73 ***
0.14
0.14 *
0.16 *
−0.09
−0.18 *
−0.13
−0.06
0.33 ***
0.37 ***
0.78 ***
0.85 ***
1.000
−0.55 ***
−0.52 ***
0.20 *
−0.01
−0.40 ***
0.10
0.13
0.10
−0.04
−0.25 **
−0.05
−0.42 ***
−0.52 ***
1.00
−0.56 ***
0.81 ***
0.74 ***
0.20 *
0.22 **
0.26 **
−0.25 ***
−0.31 ***
−0.20 *
−0.15
0.59 ***
0.48 ***
0.89 ***
0.94 ***
MBD
1.00
−0.98 *** 1.00
fMBD
−0.94 *** 0.93 ***
WMBD
PD
0.60 *** –0.42 ***
Total C
−0.81 *** 0.78 ***
oxOM
−0.78 *** 0.75 ***
Clay
−0.05
0.11
FSI
−0.12
0.13
MSI
−0.30 *** 0.24 **
FS
0.20 * −0.20 *
MS
0.21 ** −0.21 *
CS
0.08
−0.06
VCS
−0.02
0.01 *
Al oxide −0.61 *** 0.63 ***
Fe oxide −0.40 *** 0.45 ***
LL
−0.85 *** 0.86 ***
PL
−0.93 *** 0.92 ***
* Significant at P < 0.05.
PL
LL
Fe oxide
Al oxide
VCS
CS
MS
FS
MSI
FSI
Clay
oxOM
Total C
PD
WMBD
fMBD
MBD
The MBD was negatively and WMBD positively related
to Al and Fe oxides (Table 3). Linear relationships for Al
oxide with MBD and WMBD were both stronger than
those for Fe oxide (Table 3), while adding an exponential
component further improved the relationships with Al
oxide (R2 = 0.53 and 0.44; data not shown). In these relatively young soils of British Columbia that have developed
since the most recent glaciation, oxides of Fe and Al are
the main cementing agents that enhance aggregate stability
(McKeague and Sprout, 1975).
Addition of Fe oxides along with organic C, liquid
limit, and sand into the prediction model used in a study
by Howard et al. (1981) improved the predictability of
MBD by 1% (R2 = 0.99). Contrary to our findings, Fe
oxide was positively related to MBD in that study. Howard
et al. (1981) used the citrate–bicarbonate–dithionite extraction method, which removed the total Fe oxide. The positive relationship observed in their study appeared to reflect
the effect of Fe on soil mass rather than on soil strength, as
soil particle density increases with Fe content. The negative
relationship between MBD and active Fe oxide (extracted
by ammonium oxalate) in our study reflected the enhanced
soil strength due to the presence of soil oxides. Consequently,
testing of active oxides along with total oxides could provide
Variable†
Soil Oxides
Table 3. Correlation coefficients for comparisons among 17 selected variables.
In our study, MBD and WMBD were better correlated
to total C than oxidizable organic matter (Table 3). It has
been reported that oxidation of the soil organic matter by
H2O2 is restricted in the presence of Fe, which tends to
stabilize soil organic matter during oxidative degradation
(Oades and Townsend, 1963). The weaker relationship of
MBD and WMBD with oxidizable organic matter than with
total C could be attributed to the presence of Fe oxides,
indicating the importance of including several soil properties in a model to increase compactability prediction. On
the other hand, Soane et al. (1972) reported lower correlations between MBD and WMBD and organic C (R = −0.72
and −0.61, respectively) compared with oxidizable organic
matter (R = −0.81 and −0.74, respectively) for 58 British
agricultural topsoils. They used sodium dichromate mixtures to test organic C, which did not completely oxidize
the organic compounds. In addition, the reaction between
soil Fe and Mn oxides with dichromate may have further
lowered the organic C (Nelson and Sommers, 1996).
Organic matter affects the compaction process in at
least two ways: (i) it increases soil resistance to compaction by enhancing the contact between soil particles (Soane
1990); and (ii) its low particle density (Redding and Devito,
2006) compared with soil mineral particles reduces the
overall particle density and therefore bulk density, especially when organic matter content is high. For our soils,
total C accounted for 30.2% of the variation in particle
density. The coefficient of determination for the relationship between fMBD and total C, where the density effect
of total C was removed, was 5% lower than that for MBD
with total C (Table 3). Therefore, we conclude that the organic
matter had the strongest effect in improving soil resistance to
the compactive force for our group of soils.
important information to determine the mechanisms by which
these materials affect compaction.
447
Table 4. Relationships among maximum bulk density (MBD) and particle
size properties obtained at 33 study sites in British Columbia (n = 147).
R2
Model
MBD = 1.72 − 0.001(medium silt)
MBD = 1.45 + 0.0006(fine sand)
0.09
0.04
P
0.000
0.021
MBD = 1.45 + 0.001(medium sand)
0.05
0.009
MBD = 1.54 − 0.03(skewness)
MBD = 1.51 − 0.005(kurtosis)
0.01
0.00
0.335
0.732
MBD = 1.27 + 0.002(fine silt) − 0.001(medium silt) +
0.001(coarse silt) + 0.002(medium sand)
0.18
0.000
study, kurtosis ranged from −1.67 to 3.46 and skewness
varied from −0.26 to 2.26, but we found no relationship
between MBD or WMBD and kurtosis or skewness (Table 4).
Smith et al. (1997) suggested that the relationship between
MBD and kurtosis would be confounded by the significant
relationship between MBD and organic matter, which may
also have occurred in our study.
Plastic and Liquid Limits
The MBD was negatively and WMBD positively related
to liquid and plastic limits (Table 3). Liquid and plastic limits have strong linear relationships with MBD (R2 = 0.72
Particle Size Distribution
and 0.87, respectively; data not shown) and WMBD (R2 = 0.78
Our results show that particle size distribution was not
and 0.89, respectively; data not shown). An exponential model
the major factor related to variations in MBD or WMBD across
provided a better relationship between MBD and plastic limits (R2
the entire range of soils we studied, even though correlation
= 0.93; data not shown). Our results also showed that if the plastic
coefficients were significant between MBD and medium silt,
limit can be determined on a sample, it is more closely related than
and between WMBD and clay, fine silt, medium silt, fine sand,
the liquid limit to MBD and WMBD.
medium sand, and coarse sand (Table 3). In previous studies,
Plastic and liquid limits integrate several soil properties
increasing clay content has either been associated with lower MBD
such as particle size distribution, organic matter content, and
(Smith et al., 1997; Nhantumbo and Cambule, 2006) or had little
clay mineralogy. Our findings are similar to those of Soane et
effect on MBD (Ball et al., 2000; Aragon et al., 2000).
al. (1972), who tested 13 properties of 58 Scottish topsoils and
Kurtosis and skewness were previously considered to be
found that MBD and WMBD were highly related to the plastic
useful parameters in predicting soil MBD (Moolman, 1981). A
and liquid limits (R = −0.80 and −0.68 for MBD, 0.74 and
low coefficient of kurtosis indicates a well-graded particle size
0.69 for WMBD). Howard et al. (1981) found that the MBD
distribution and is expected to lead to a higher MBD, while a
of California forest and rangeland soils was significantly corlow absolute value of the skewness coefficient indicates high
related to the liquid limit (R = −0.96) but they did not report
symmetry of the distribution curve and higher MBD. In our
the plastic limit. Ball et al. (2000) reported that the liquid
limit accounted for 43 and 48% of the variation in
Table 5. Principal component analysis loadings for the first three components MBD and WMBD, respectively, of British soils, while
the relationship between MBD or WMBD and the
of individual variables (n = 147).
plastic limit was lower (R2 = 0.29 and 0.36, respecLoadings‡
Variable†
tively). Relative to our data, the correlation coeffiComponent 1§ Component 2¶ Component 3#
cient between MBD and liquid limit was higher in a
MBD
−0.15
0.00
−0.33
study by Howard et al. (1981) and lower in Ball et al.
0.14
0.05
fMBD
0.32
0.09
0.01
WMBD
0.34
(2000). The former study tested 14 Californian soils
Particle density
−0.19
−0.14
0.17
predominated by loam texture, while the latter evaluTotal C
0.11
0.00
0.30
ated 146 British agricultural soil samples with a wide
Oxidizable organic matter
0.04
0.01
0.29
variation
in texture. Differences in the soil textures
Clay
0.12
0.16
−0.38
might have accounted for the difference in correlation
Fine silt
0.14
0.06
−0.35
Medium silt
0.13
−0.04
−0.28
coefficients among the above studies. In the study by
Coarse silt
−0.03
−0.01
−0.48
Ball et al. (2000), lower correlation coefficients may
Very fine sand
−0.08
0.20
−0.41
have been observed because some nonplastic soils had
Fine sand
−0.16
0.19
0.30
missing
values filled by a statistical tool before running
Medium sand
−0.16
0.29
0.29
the correlation analysis.
Coarse sand
−0.13
0.33
0.26
Very coarse sand
Kurtosis
Skewness
Al oxide
Fe oxide
Mn oxide
Si oxide
Plastic limit
Liquid limit
−0.05
0.07
0.08
0.17
0.20
0.16
0.08
0.33
0.33
0.27
−0.20
−0.20
0.29
−0.04
0.14
0.24
0.08
−0.01
0.15
0.36
0.32
0.05
0.01
0.13
0.01
−0.04
0.08
† MBD, maximum bulk density; fMBD, porosity at MBD; WMBD, water content at
which MBD was achieved.
‡ Values >0.25 are italicized.
§ Accounted for 34% of variation.
¶ Accounted for 19% of variation.
# Accounted for 14% of variation.
448
Predicting Maximum Bulk Density by a
Set of Soil Properties
Principal component analysis (a multivariate
analysis tool to examine relationships among several
quantitative variables in a data set) showed that the
first three components accounted for 67% of the variation in the data set (Table 5). The first component
mainly explained MBD, fMBD, and WMBD and soil
properties like liquid and plastic limit, total C, and
oxidizable organic matter. The second and third components mainly explained soil texture and Al oxides.
This indicated that soil organic matter and liquid and
plastic limits had the greatest impact on MBD, fMBD,
SSSAJ: Volume 72: Number 2 • March–April 2008
and WMBD, while particle density, Al and Fe oxides, and some
of the particle size classes were of secondary importance.
Principal component analysis allows a reduction in the
number of variables used in regression analysis because it identifies factors whose effects are independent of one another.
Multiple regression analysis was performed to find the best
combination of soil properties that would explain variation in
MBD. Because it was not possible to obtain the plastic limit
for >20% of the samples, and considering that the plastic limit
was previously shown (Ball et al., 2000) to be an important
factor affecting MBD, we performed separate multiple regression analyses on soil groups based on their plasticity as shown
in Fig. 3. Soils with high plasticity were characterized by either
high clay content (up to 700 g kg−1) or high total C (up to
77 g kg−1). Moderately plastic soils had lower contents of clay
(up to 560 g kg−1) and total C (up to 57 g kg−1), and made
up the largest group of soils in our study. Nonplastic soils had
the lowest clay content (up to 170 g kg−1) and variable total
C content (4–63 g kg−1). Generally, it is difficult to determine
the plastic limit on the very coarse-textured soils that cover
some areas of British Columbia.
We were able to predict the MBD of British Columbia forest soils by combining several soil properties (Table 6). When
all samples were included in the regression analysis, the liquid
limit was the most highly correlated property in explaining
MBD among all soil properties included in this study. The liquid limit, in combination with clay content, explained >80%
of the variation in MBD. When oxidizable organic matter and
Al oxide were added to the liquid limit and clay content, predictability of MBD improved by 8% (Table 6).
When samples were grouped according to their plasticity, fewer variables were needed in the multiple regressions to
explain comparable amounts of variation in MBD, compared
Fig. 3. Plasticity of soils from the study areas, showing highly plastic soils
with liquid limit >0.50 and soils with moderate and low plasticity
(liquid limit <0.50) as plotted on the Casagrande chart. The A-line
represents the division between clays (plot on or above the A-line)
and silts (plot below the A-line); the U-line refers to the upper limit.
with the entire sample set. Multiple regressions for the nonplastic soils explained the most variation, while those for the
highly plastic soils explained the least, and soils with low and
moderate plasticity were intermediate (Table 6). For the nonplastic soils (i.e., those with low clay content), liquid limit and
Al oxide were the two most important properties in predicting
MBD (R2 = 0.96). In the moderately plastic group, the plastic
limit and oxidizable organic matter were the first two properties
entered into the regression to predict MBD (R2 = 0.89). For
highly plastic soils (i.e., those with high clay and organic matter contents), total C and the plastic limit explained 87% of
the variation in MBD.
Table 6. Regression constants and correlation coefficients for relationships between maximum bulk density (MBD) as the dependent
variable and selected soil properties as the independent variable.
Dependent variable
Overall (n = 144)
Intercept Liquid limit
MBD
R2 ‡
Independent variable†
2.07
2.06
2.06
2.02
−2.11
−1.61
−1.29
−1.35
Clay
0.0006
0.0006
0.0004
0.0005
Oxidizable organic
matter
−0.005
−0.005
−0.005
Al oxide
−0.17
−0.16
Very coarse
sand
0.83
0.88
0.91
0.92
0.0005
Nonplastic (n = 29)
Intercept Liquid limit
2.09
−1.79
2.07
−1.89
1.98
−1.61
Moderately plastic (n = 99)
MBD
Intercept Plastic limit
MBD
2.21
2.26
2.24
2.28
2.27
−2.24
−2.16
−1.83
−1.79
−1.62
Al oxide
−0.14
−0.12
−0.11
Oxidizable organic
matter
−0.004
−0.004
−0.003
−0.003
−0.003
Very coarse sand
Total C
Clay
0.0005
0.0006
−0.003
0.0006
Medium silt
Total C
Fine silt
−0.0003
−0.0004
−0.0004
−0.0005
−0.002
−0.003
−0.003
−0.0005
−0.0005
0.96
0.97
0.98
Al oxide
−0.18
0.89
0.90
0.91
0.92
0.92
Highly plastic (n = 16)
Intercept
Total C
Plastic limit
MBD
1.72
−0.004
−0.82
0.87
†Liquid and plastic limits (kg kg−1); Al and Fe oxide (%); oxidizable organic matter, total C, clay, medium silt, fine silt, fine sand, and very coarse sand (g kg−1).
‡ Significant at P < 0.001.
SSSAJ: Volume 72: Number 2 • March–April 2008
449
the positive effect we observed was
opposite to that reported by Smith et
al. (1997). Smith et al. (1997) found
that in the lower range of clay + silt
(0–400 g kg−1), the effectiveness of
clay + silt and total C appeared to
offset each other in the compaction
test; only in the higher range of clay
+ silt (400–1000 g kg−1) did clay +
silt enhance the effect of total C in
reducing MBD. Hence, the importance of these two properties cannot be compared directly in their
study. As total C was positively correlated to clay + silt (R2 = 0.33) in
their study, the strong effect of clay
on MBD may be just a covarying
result of organic matter on MBD.
For the highly plastic group in our
study, there was also a close correlation between clay and total C, which
was not apparent in the other groupings (Fig. 5), but the relationship we
observed showed total C content
declining with increasing clay content, opposite to what is commonly
expected for a range of soils. Our
highly plastic soils group appeared to
contain a mix of two subgroups of
Fig. 4. Relationships among maximum bulk density (MBD) and clay + silt (fsi = fine silt; msi = mesoils: (i) those with very high clay condium silt) for (a) all samples, (b) nonplastic samples, (c) moderate and low plastic samples, and
tent but low or intermediate organic
(d) highly plastic samples. ***Significant at P < 0.001.
matter; and (ii) those with very high
For the majority of our soils, oxidizable organic matter
organic matter content but low clay content. This may clarify why
was preferred to total C when predicting MBD, illustrating
multiple regressions explained the smallest amount of variation for
the importance of quality-related (i.e., oxidizable organic matthe highly plastic group.
ter) rather than quantity-related (i.e., total C) soil organic matBecause compaction is a dynamic process, the surface area,
ter in compaction studies. In a study by Howard et al. (1981),
contacting points, and surface charge tend to be more impororganic C was the most important variable in predicting MBD,
tant than single particles. Soil properties that more directly repbut they did not determine the active oxides and subgroups of
resent the above mechanisms should give a good description
sand and silt size fractions, hence the importance of oxidizable
of the compaction process. Plastic and liquid limits have been
organic matter was not known. Oxidizable organic matter was
proven powerful (R2 > 0.90) in estimating external surface area
the second most important variable in predicting MBD in a
(Hammel et al., 1983); on the other hand, oxalate-extractable
study by Ball et al. (2000).
oxides reflect the charge condition of particle surfaces. Organic
In our study, both organic matter (either total C or oxidizmatter may also be more important than particle size distribuable organic matter) and oxides were important for the prediction where living and dead roots provide a filamentous network,
tion of MBD (Table 6). In all groups, organic matter (i.e., total
which resists compactive loads, and highly humified material
C) showed a strong relationship in decreasing MBD (R2 = 0.59–
increases the stability of aggregates (Soane, 1990).
0.72), which is similar to the results of Smith et al. (1997), who
Proposed Method for Using Maximum Bulk Density
found a strong negative relationship between MBD and total
2
as a Reference in Forest Soil Compaction Studies
C (R = 0.88), and also to Aragon et al. (2000), who showed a
high dependence of MBD on organic C. Including particle size
To use MBD as a reference value in soil compaction studdistribution further improved the prediction. Even though clay
ies, a method for obtaining the best estimate of MBD across a
came second in predicting MBD for the “overall” group, parvariable site is required. The standard approach to determining
ticle size components usually ranked third or lower in the preMBD using the Proctor test relies on collecting a 10-L sample
diction. Unlike Smith et al. (1997), who found a strong relafrom the site and carrying out the laboratory test. On typical
tionship between MBD and clay + silt (R2 = 0.63), there was
forestry sites, such a method may be impractical because site
no relationship between MBD and clay + silt in our nonplastic
variability makes it difficult to identify the “typical” condition
and moderately plastic groups. Only in the highly plastic group
that will best represent conditions throughout the site. A better
did clay + silt show a high correlation with MBD (Fig. 4), but
approach would be to collect samples from each variant of the
450
SSSAJ: Volume 72: Number 2 • March–April 2008
soil conditions, but this too may become
unwieldy because of the large number of
samples required. Generally, bulk density
sampling requires high numbers of samples
to account for the natural variation on forestry field sites (Courtin et al., 1983; PageDumroese et al., 1999). The method we
propose takes advantage of the strong relationships we have observed between MBD
and soil properties that are relatively easy to
measure. The method involves four steps.
1.
Determine the relationships
between MBD and properties for
soils typical of the study area. As
we, and others have shown, MBD
can be predicted with reasonable
accuracy from a relatively small
number of properties, but the
best properties for prediction
may be different for different
groups of soils. The properties
to use in the prediction of MBD
can be selected by stratification of
samples from a larger data set, as
we have described. We stratified
our sample set based on plasticity,
but other approaches could be
applicable in a particular study.
Fig. 5. Relationships between clay and total C for (a) all samples, (b) nonplastic samples, (c) moderate and low plastic samples, and (d) highly plastic samples. ***Significant at P < 0.001.
2. Collect bulk density samples from
the field sites.
3. Carry out laboratory analyses on the bulk density
samples to provide data to be used in multiple
regression analysis as we have described here. The
analysis will produce a “predicted MBD” for each soil
sample that will account for the fine-scale variation
in soil properties typical of forestry sites. It may be
possible to carry out the analysis for different variables
than we have described, depending on the needs of
the study and the resources available. For example, in
the nonplastic soils of our study, the use of total C
and Al oxide explained a large amount of variation in
MBD (R2 = 0.88), although not as much as the liquid
limit and Al oxide (R2 = 0.96).
4. Develop empirical relationships between field bulk density,
MBD, and tree growth. We are conducting further
investigations to test the applicability of relative measure
of bulk density (i.e., field bulk density/MBD) for
compaction studies on forest soils in British Columbia
that have been described previously (Carter, 1990; da
Silva and Kay, 1997).
CONCLUSIONS
The significance levels of single soil properties in predicting MBD were in the order of liquid and plastic limits, organic
matter, and oxalate-extractable oxides, while particle size distribution alone accounted for very little variation. In the mulSSSAJ: Volume 72: Number 2 • March–April 2008
tiple regression analysis for the entire sample set, liquid limit
and clay were related to MBD. Inclusion of organic matter,
Al oxides, and other components of the particle size distribution (e.g., very coarse sand) further improved the prediction of
MBD. Stratification of the sample set by plasticity allowed substantially improved prediction of MBD using multiple regression analysis. The best predictions were obtained for nonplastic
soils, while multiple regression explained the least amount of
variation for highly plastic samples. Porosity at MBD may be
useful for studies relating plant growth to soil physical condition. On the other hand, use of MBD may be preferred over
fMBD for evaluating soil conditions where a reference value for
soil bulk density is required.
Currently, only bulk density is used widely as a parameter
to assess the compaction state of a soil. We have described a
method to predict MBD from readily measured soil properties that could enable more effective means of providing reference values for compaction studies. This would be particularly
beneficial where these attributes exhibit high point-to-point
variation, such as in British Columbia’s forest soils. Prediction
would involve first determining the plasticity for a soil sample,
then using the appropriate equation to determine MBD.
ACKNOWLEDGMENTS
This work was supported by the Natural Sciences and Engineering Research
Council (NSERC) of Canada and the Forest Investment Account—Forest
Science Program. Technical assistance of Mike Curran, Bob Maxwell, Lesley
Dampier, Francois Teste, Shakedur Rahman, Chris Blurton, Korey Green,
Sara Harrison, Matthew Plotnikoff, Allison Tremain, Rick Trowbridge,
451
and Will Arnup in the laboratory and with sample collection is greatly
appreciated. We also acknowledge the helpful suggestions from Dr. Les
Lavkulich during preparation of this manuscript.
REFERENCES
Aragon, A., M.G. Garcia, R.R. Filgueira, and Ya.A. Pachepsky. 2000. Maximum
compactibility of Argentina soils from the Proctor test: The relationship
with organic matter and water content. Soil Tillage Res. 56:197–204.
ASTM. 2000. Design D698–00a. Standard test methods for laboratory compaction
characteristics of soil using standard effort (12,400 ft-lb/ft3 (600 kN-m/m3)).
ASTM, West Conshohocken, PA.
Ball, B.C., D.J. Campbell, and E.A. Hunter. 2000. Soil compactibility in relation to
physical and organic properties at 156 sites in UK. Soil Tillage Res. 57:83–91.
Ball, B.C., R.W. Lang, M.F. O’Sullivan, and M.F. Franklin. 1989. Cultivation
and nitrogen requirements for continuous winter barley on a Gleysol
and a Cambisol. Soil Tillage Res. 13:333–352.
Carter, M.R. 1990. Relative measures of soil bulk density to characterize compaction
in tillage studies on fine sandy loams. Can. J. Soil Sci. 70:425–433.
Courtin, P., M.C. Feller, and K. Klinka. 1983. Lateral variability in some
properties of disturbed forest soils in southwestern British Columbia.
Can. J. Soil Sci. 63:529–539.
da Silva, A.P., and B.D. Kay. 1997. Estimating the least limiting water range of
soils from properties and management. Soil Sci. Soc. Am. J. 61:877–883.
da Silva, A.P., B.D. Kay, and E. Perfect. 1994. Characterization of the least
limiting water range of soils. Soil Sci. Soc. Am. J. 58:1775–1781.
Day, P.R. 1965. Particle fractionation and particle size analysis. p. 545–567. In
C.A. Black (ed.) Methods of soil analysis. Part 1. Agron. Monogr. 9. ASA
and SSSA, Madison, WI.
Donkin, M.J. 1991. Loss-on-ignition as an estimator of soil organic carbon in
A-horizon forestry soils. Commun. Soil Sci. Plant Anal. 22:233–241.
Dorel, M., J. Roger-Estrade, H. Manichon, and B. Delvaux. 2000. Porosity
and soil water properties of Caribbean volcanic ash soils. Soil Use
Manage. 16:133–140.
FAO. 1998. World reference base for soil resources. World Soil Resour. Rep.
84. FAO, Rome.
Flint, A.L., and L.E. Flint. 2002. Particle density. p. 229–240. In J.H. Dane
and G.C. Topp (ed.) Methods of soil analysis. Part 4. SSSA Book Ser. 5.
SSSA, Madison, WI.
Froehlich, H.A., D.W.R. Miles, and R.W. Robbins. 1985. Soil bulk-density
recovery on compacted skid trails in central Idaho. Soil Sci. Soc. Am. J.
49:1015–1017.
Gee, G., and D. Or. 2002. Particle-size analysis. p. 255–293. In J.H. Dane and
G.C. Topp (ed.) Methods of soil analysis. Part 4. Physical methods. SSSA
Book Ser. 5. SSSA, Madison, WI.
Greacen, E.L., and R. Sands. 1980. Compaction of forest soils: A review. Aust.
J. Soil Res. 18:163–189.
Hakansson, I., and J. Lipiec. 2000. A review of the usefulness of relative bulk density
values in studies of soil structure and compaction. Soil Tillage Res. 53:71–85.
Hammel, J.E., M.E. Sumner, and J. Burema. 1983. Atterberg limits as indices
of external surface areas of soil. Soil Sci. Soc. Am. J. 47:1054–1056.
Howard, R.F., M.J. Singer, and G.A. Frantz. 1981. Effects of soil properties,
water content, and compactive effort on the compaction of selected
California forest and range soils. Soil Sci. Soc. Am. J. 45:231–236.
Joosse, P.J., and R.A. McBride. 2003. Assessing physical quality of plastic
soils of differing mineralogy and pre-stress history using mechanical
parameters: I. Saturated compression tests. Can. J. Soil Sci. 83:45–63.
Krzic, M., C.E. Bulmer, F. Teste, L. Dampier, and S. Rahman. 2004. Soil
properties influencing compactability of forest soils in British Columbia.
Can. J. Soil Sci. 84:219–226.
Larson, W.E., S.C. Gupta, and R.A. Useche. 1980. Compression of agricultural
soils from eight soil orders. Soil Sci. Soc. Am. J. 44:450–457.
Lipiec, J., I. Hakansson, S. Tarkiewicz, and J. Kossowski. 1991. Soil physical
properties and growth of spring barley as related to the degree of
compactness of two soils. Soil Tillage Res. 19:307–317.
Lipiec, J., and R. Hatano. 2003. Quantification of compaction effects on soil
physical properties and crop growth. Geoderma 116:107–136.
Loeppert, R.L., and W.P. Inskeep. 1996. Iron. p. 639–664. In D.L. Sparks et al. (ed.)
Methods of soil analysis. Part 3. SSSA Book Ser. 5. SSSA, Madison, WI.
452
McBride, R.A. 2002. Atterberg limits. p. 519–527. In J.H. Dane and G.C. Topp
(ed.) Methods of soil analysis. Part 4. SSSA Book Ser. 5. SSSA, Madison, WI.
McKeague, J.A., and P.N. Sprout. 1975. Cemented subsoils (duric horizons)
in some soils of British Columbia. Can. J. Soil Sci. 55:189–203.
Meidinger, D.V., and J. Pojar. 1991. Ecosystems of British Columbia. Spec.
Rep. Ser. 6. British Columbia Ministry of Forests, Victoria.
Miller, R.E., W. Scott, and J.W. Hazard. 1996. Soil compaction and conifer
growth after tractor yarding at three coastal Washington locations. Can.
J. For. Res. 26:225–236.
Moolman, J.H. 1981. Soil textural properties influencing compactibility.
Agrochemophysica 13:13–19.
Nelson, D.S., and L.E. Sommers. 1996. Total carbon, organic carbon, and organic
matter. p. 961–1010. In D.L. Sparks et al. (ed.) Methods of soil analysis. Part
3. SSSA Book Ser. 5. SSSA, Madison, WI.
Nhantumbo, A.B.J.C., and A.H. Cambule. 2006. Bulk density test as a function
of texture for agricultural soils in Maputo province of Mozambique. Soil
Tillage Res. 87:231–239.
Oades, J.M., and W.N. Townsend. 1963. The influence of iron on the stability
of soil organic matter during peroxidation. J. Soil Sci. 14:134–143.
Padmanabhan, E., and A.R. Mermut. 1995. The problem of expressing the
specific surface areas of clay fractions. Clays Clay Miner. 43:237–245.
Page-Dumroese, D.S., A.E. Harvey, M.F. Jurgensen, and M.P. Amaranthus. 1998.
Impacts of soil compaction and tree stump removal on soil properties and
outplanted seedlings in northern Idaho, USA. Can. J. Soil Sci. 78:29–34.
Page-Dumroese, D.S., M.F. Jurgensen, R.E. Brown, and G.D. Mrosz. 1999.
Comparison of methods for determining bulk densities of rocky forest
soils. Soil Sci. Soc. Am. J. 63:379–383.
Powers, R.F. 2006. Long-term soil productivity: Genesis of the concept and
principles behind the program. Can. J. For. Res. 36:519–528.
Proctor, R.R. 1933. Fundamental principles of soil compaction. Description
of field and laboratory methods. Eng. News Rec. 111:286–289.
Redding, T.E., and K.J. Devito. 2006. Particle densities of wetland soils in
northern Alberta, Canada. Can. J. Soil Sci. 86:57–60.
Ruhlmann, J., M. Korschens, and J. Graefe. 2006. A new approach to calculate
the particle density of soils considering properties of the soil organic
matter and the mineral matrix. Geoderma 130:272–283.
SAS Institute. 1990. SAS/STAT user’s guide. Version 6. SAS Inst., Cary, NC.
Sheldrick, B.H., and C. Wang. 1993. Particle size distribution. p. 499–512.
In M.R. Carter (ed.) Soil sampling and methods of analysis. Lewis Publ.,
Boca Raton, FL.
Smith, C.W., M.A. Johnston, and S. Lorentz. 1997. Assessing the compaction
susceptibility of South African forestry soils: II. Soil properties affecting
compactibility and compressibility. Soil Tillage Res. 43:335–354.
Soane, B.D. 1990. The role of organic matter in soil compactibility: A review
of some practical aspects. Soil Tillage Res. 16:179–201.
Soane, B.D., D.J. Campbell, and S.M. Herkes. 1972. The characterization of
some Scottish arable topsoils by agricultural and engineering methods. J.
Soil Sci. 23:93–104.
Soil Classification Working Group. 1998. The Canadian system of soil
classification. 3rd ed. Agric. and Agri-Food Can. Publ. 1646. Natl. Res.
Counc. of Canada, Ottawa, ON.
Soil Survey Staff. 2006. Keys to Soil Taxonomy. 10th ed. NRCS, Washington, DC.
Topp, G.C., W.D. Reynolds, F.J. Cook, J.M. Cook, and M.R. Carter. 1997. Physical
attributes of soil quality. p. 21–58. In E.G. Gregorich and M.R. Carter (ed.)
Soil quality for crop production and ecosystem health. Elsevier, Amsterdam.
Webster, R., and M.A. Oliver. 1990. Statistical methods in soil and land
resources survey. Oxford Univ. Press, Oxford, UK.
Wert, S., and B.R. Thomas. 1981. Effects of skid roads on diameter, height,
and volume growth in Douglas-fir. Soil Sci. Soc. Am. J. 45:629–632.
Yee, C.S., and R.D. Harr. 1977. Effect of wetting mode on shear strength of
two aggregated soils in the Oregon coast ranges. USDA For. Serv. Res.
Note PNW-303. Pac. Northw. For. Range Exp. Stn., Portland, OR.
Zhang, H., K.H. Hartge, and H. Ringe. 1997. Effectiveness of organic matter
incorporation in reducing soil compactibility. Soil Sci. Soc. Am. J.
61:239–245.
Zou, C., C. Penfold, R. Sands, R.K. Misra, and I. Hudson. 2001. Effects of
air-filled porosity, soil matric potential and soil strength on primary root
growth of radiate pine seedlings. Plant Soil 236:105–115.
SSSAJ: Volume 72: Number 2 • March–April 2008